WHITE PAPER

Analytics in Healthcare
How the healthcare industry will uncover the real value of electronic
medical records and the emerging electronic health record (EHR) initiative
 Analytics in Healthcare

Table of Contents

Executive summary..........................................................................................1
The paradoxical state of healthcare technology.............................................2
Promising trends for adoption of electronic medical records........................4
EMR and EHR – There is a difference...........................................................5
Will EMRs and EHRs be the blockbuster prescription?...................................6
SAS solutions for healthcare providers...........................................................8
The data-based evolution of medicine..........................................................10

i
 Analytics in Healthcare

ii
 Analytics in Healthcare

Executive summary

Electronic medical records (EMRs) and electronic health records (EHRs) hold
great promise to improve the efficiency and quality of patient care while reducing
cost and errors. The development of standard data models and a standard
lexicon for coding patient care – coupled with government endorsement for a
national healthcare information infrastructure – is accelerating the transition to
paperless practice.

For all its promise, though, there are drawbacks. Electronic record-keeping
systems are expensive. Physicians are hesitant to change their work habits.
Until everybody adopts electronic methods, the truly paperless healthcare
environment remains a pipe dream.

Even when that point is reached, EMR and EHR systems alone will not offer up
all of the benefits that they should. These systems will automate transactions,
support research and make records more accessible, but will they answer
elemental questions about the quality of those patient interactions?

• What treatment regimen yields the best outcomes for patients with this
genetic profile?

• What insidious drug interactions are we likely to see in a patient with these
risk factors?

• How well does this combination of therapies help patients undergoing

this procedure?

• What protocol produces the best rehabilitation results for this target
population?

To answer questions such as these, clinicians and researchers need business

intelligence (BI). Their administrative staff have been using BI for years to run the
business better. Now, with new sources of digital data about patients and their
clinical experiences, BI is being applied to improve not just medical practices, but
the practice of medicine.

The near-term evolution is already exciting. Traditionally, medicine focused

on diagnosing and treating existing conditions, with limited data links across
the continuum of care. In the emerging stage of data-powered healthcare,
predictive analytics are being applied to reduce errors and to improve outcomes
– incorporating new data sources such as genomic data and digital diagnostic
images. The logical extension of this evolution is personalized medicine – holistic,
pre-symptomatic treatment focused on preventing conditions rather than treating
them after the fact.

1
 Analytics in Healthcare

The common denominator for these evolutionary stages is data – the ability to gather,
cleanse and analyze extremely diverse data, without having to be a statistics guru.

The good news is that the technology is here today.

So, bring on the evolution and press EMR/EHR systems to deliver the real return on
investment that they can.

The paradoxical state of healthcare technology

No question about it, modern-day health care is a technological marvel.

Microscopic laser pulses reshape the human cornea to restore perfect vision.
Magnetic resonance imagery and computerized axial tomography show us intimate
details of living tissue, even the human brain. Robotic systems with tiny cameras and
instruments perform delicate microsurgeries. Sound waves create three-dimensional
images of unborn babies. Artificial hearts made of titanium and plastic sustain the
lives of humans after their natural hearts have failed.

But darned if you can find the chart for a new patient being transferred from the
satellite clinic… Or decipher the scrawled prescription… Or get a complete history
on the patient who can’t remember.

“It is astonishing that, in a time when computers enable businesses to manage and
locate everything from dry cleaning to used trucks, few healthcare providers are
able to access and track the medical records of patients across the continuum of
care,” wrote Charlene Marietti, editor of Healthcare Informatics1. “Most large acute
care providers are aggressively tackling the electronic medical record deficiency, but
providers in thousands of solo and small office practices, where the majority of care
is delivered, show little progress toward that goal.”

An electronic medical record (EMR) is a computer-based record of patient care.

EMRs have the potential to enhance the efficiency of every aspect of the patient
experience, from scheduling appointments to managing patient care to reporting
and billing. A computer-based record has the advantage of being accessible from
multiple locations via a secure network. Providers can have ready access to medical
histories, lab test results, digitized diagnostic imagery and reference databases –
wherever they are.

1 Marietti, Charlene. “Mountains to Climb: Cost isn’t the biggest obstacle on the way to EMRs,”
Healthcare Informatics (October 2004)

2
 Analytics in Healthcare

No one disputes the potential value of standardized EMRs. Study after study reveals
that electronic medical records can improve care and reduce costs. The projections
are enough to make a hospital administrator salivate: promises to eliminate more
than 2 million adverse drug events and 190,000 unnecessary hospitalizations a
year… to reduce medication processing time by 68 percent and problem medication
orders by 58 percent… to trim 10 percent from the nation’s $1.7 trillion a year
healthcare bill… or as much as $400 billion saved per year, depending on whom
you ask.

Why, then, is 60 percent of hospital “paperwork” still actually on paper? Why is only
15 percent of clinical data stored in digital format? Why does North America lag so
far behind other developed countries in the adoption of electronic records?

“Although multiple stakeholders, including the government, insurers and even

consumers, are pushing this change engine, the disadvantages continue to outweigh
the advantages for the majority of private practice physicians,” Marietti wrote. “It isn’t
a question of technology. Physicians aren’t Luddites. They own and operate some of
the most sophisticated equipment and devices in the world.”

But even now, a half-decade into the new millennium, the deterrents seem to
outweigh the benefits:
The eight core elements of
Change is expensive. An EMR system for a large hospital or healthcare network the IOM data model
can cost $10 million or more. A system for a smaller clinic or group practice can cost
$10,000 to $20,000 per physician. And that’s just for starters. Ongoing operating 1. Health information and data
costs can add another 25 percent per year.
2. Results management
Any change, even positive change, is a burden. “Many physicians recognize 3. Order entry/order management
the potential of EMRs but remain sidelined by the extreme difficulties they foresee
on every path to adoption,” Marietti wrote. Many physicians simply prefer paper- 4. Decision support
based processes to electronic ones. Even if they don’t have a bias, they know how
disruptive it will be to change their work habits and move thousands of pages of
5. Electronic communication
paper records into electronic files. and connectivity

6. Patient support
The paperless practice is ultimately a long-term vision. “Even physicians who
do have EMRs cannot fully realize their return on investment until the majority of 7. Administrative processes
practices are computer-based,” Marietti noted. A team that adopts an EMR may
enjoy internal efficiencies, but until everybody uses the technology, staff will still have 8. Reporting and population
to grapple with paper coming in from clinical departments, specialists and other health management
clinics. They may even have to add staff to scan and enter all of that paper-based
information into new computer systems.

So until very recently, EMR remained a good idea bogged down by lack of standards
and market acceptance.

3
 Analytics in Healthcare

Promising trends for adoption of electronic

medical records

The last few years have seen several trends that are accelerating the adoption and
value of EMR:

Government endorsement. In July 2004, the U.S. Department of Health and

Human Services (HHS) announced a 10-year plan to create a new National Health
Information Infrastructure (NHII), including an EHR for every American and a new
network to link health records nationwide. According to then HHS Secretary Tommy
Thompson, the initiative “will provide a quantum leap in patient power, doctor power
and effective health care.”

A standard health record model. HHS commissioned the Institute of Medicine (IOM)
to design a standard EHR model for the healthcare industry – working with Health
Level 7 (HL7), an American National Standards Institute (ANSI)-affiliated standards
organization. This initiative gave rise to the EHR Collaborative, a broad-based
consortium of public and privatesector healthcare organizations, which generated
data models for IOM approval. A national EHR standard – as part of the NHII – will
enable effective data sharing among all stakeholders.

Standard language. A standard model needs a uniform language. That goal got a
huge boost when the National Institutes of Health National Library of Medicine signed
a five-year contract with the College of American Pathologists to license SNOMED-
CT (Systematized Nomenclature of Medicine – Clinical Terms). The contract provides
universal access to a machine-readable, clinically rich lexicon for standardized
coding.

Better mobile computing options. Some physicians still shun computers, but more
of them than ever are embracing handhelds, tablet PCs and laptops. That’s no
surprise, since every year mobile devices get faster, cheaper and easier to use. EMR
software is getting more mature and user-friendly, with interfaces that accommodate
more diverse practice patterns. Wireless networks are growing ever more reliable and
affordable, with better coverage. The critical pieces of the EMR technology puzzle
are falling into place.

Growing market acceptance. In a recent member survey by the American

Academy of Family Physicians, more than 80 percent of respondents said they had
investigated acquiring an EMR system. Granted, most of them still don’t have the
technology, but their willingness even to window-shop is encouraging. Perhaps
the trend represents a generational shift, as younger physicians emerge from
tech-intensive medical school programs. Maybe it’s the natural result of improved
technology options. Or perhaps it’s simply a logical response to fiscal pressures;
healthcare providers are tasked to do more with less and technology is proving
to help.

4
 Analytics in Healthcare

Early signs of positive results. Clinicians are moved to action when they see the
difference these systems make in patient care delivery. EMR ultimately is about
improving the practice of medicine and enhancing the workflow of physicians and
their interactions with patients, administrators and researchers. As providers start to
see the benefits, we will see a groundswell in adoption of EMR technology.

EMR and EHR – There is a difference

The two terms are often used interchangeably, but they do have different
meanings:

• An electronic medical record (EMR) is a computer-based medical record of a

patient’s activity within a provider organization. This record may incorporate
outside data, such as that from labs and referring specialists, but the EMR is
generally focused on activity within a single provider. The record is typically
owned and controlled by the provider organization and used to support
transactions and access across that organization.

• An electronic health record (EHR) operates at a much higher level. An EHR

contains a lifetime record of a patient’s medical history across many provider
organizations and technology platforms. In the vision set forth by HHS of
a national EHR data store, this record is owned by the patient, who grants
access to providers.

Provider-centric EMRs feed into patient-centric EHRs. However, the real

promise of these records will be seen when EMRs are integrated with practice-
management systems and EHRs are integrated with research systems to
support evidence-based healthcare delivery. More on that later.

5
 Analytics in Healthcare

Will EMRs and EHRs be the blockbuster prescription?

Or will they just automate transactions and generate more data?

The widespread adoption of EMR systems (and in turn, a national EHR database)
will no doubt be a boon to efficiency and effectiveness. Healthcare organizations
can save time, reduce costs and improve processes by automating patient care
transactions, such as appointment scheduling, medication orders, lab tests
and billing.

However, these benefits only begin to scratch the surface. Transactional EMR
systems will only go so far in delivering ROI. Alone, these systems will not provide the
insights that would enhance the quality of patient care and the practice of medicine in
general.

In short, there may be more data than ever, but will it truly be answering stakeholders’
needs?
■ The answers lie within mountains
• Clinical investigators and research administrators need to track outcomes with
more reliability and thoroughness than general clinical practice, while satisfying
of clinical, research and practice
institutional review boards and the new NHII framework. management data. Unfortunately,
this data lies buried in a labyrinth
• Regulators and financial backers need assurances that studies have adhered to
of disparate systems and databases
rigorous methodology, appropriate legal frameworks and guidelines for safety
and efficacy. that are neither integrated nor
fully utilized.
• Clinical managers wonder, “How can I accurately investigate patient and clinic
data over time to improve the overall quality of patient care and make the best
operational decisions?”

• Public health officials and pharmaceutical companies need to know how clinical
research results relate to large populations and other studies in the entire
research domain.

• Clinicians are asking, “Can I have faith in the research reports that I read and the
implications they hold for modifying my treatment plans?”

• The public is asking, “Are decisions about our medical care based on the best
possible evidence and advocacy?”

The current mode of operation in many healthcare organizations doesn’t provide very
effective answers to these questions, or provides them at very high cost and with
dubious accuracy. Even as standards start to take hold at the national level, there are
practical, local impediments that make it difficult to gain the necessary insights:

6
 Analytics in Healthcare

• Different types of data. Even when clinical data is available in digital form, it is
usually formatted for billing purposes rather than for analysis. Furthermore, there
are several entirely different categories of data to deal with: EMR data that relates
directly to patient care, aggregated data about organizational performance and
resource utilization, statistically derived data for planning and decision support and
comparative data for research and outcome assessment.

Some of this data is episodic and patient-focused, some is cumulative over

time, entities and populations. Data elements and coding strategies vary among
specialties, such as MedDRA for recording adverse events and ICD-9CM for
coding diagnoses for reimbursement.

It’s a challenge to reconcile and cleanse all of this incompatible data, much less
reap useful intelligence from it.

• Incompatible platforms. Traditional healthcare information systems typically

reflect a process- and patient-oriented view of the business. The architecture uses
a host of independent systems on different platforms, which share information in
a limited way, if at all. As growth, mergers and acquisitions reshape information
networks, it’s common to see multiple, incompatible platforms even within a single
functional area.

To realize the real value potential of EMR/EHR, the industry needs a more holistic
approach – the ability to combine EMR data with other types of data (lab, financial,
operational, research, etc.) and analyze it to reveal the hidden knowledge it contains
about trends and opportunities. This calls for “business intelligence” and data
analysis.

“The use of business-intelligence and data-analysis software in the healthcare

industry isn’t new,” said Rick Whiting in a special Healthcare Enterprise edition
of Information Week2. “Such tools have been a mainstay on the financial and
administrative side for a decade or more.”

Whiting continued: “Until recently, however, business-intelligence technology hasn’t

played a big role in delivering care. That’s changing as healthcare companies build
data repositories with information culled from operational systems used to record
patient admissions and discharges, bill patients and insurers, order laboratory and
radiological tests and dispense medications. The data can be analyzed to judge
alternative treatments such as heart bypass surgery versus angioplasty by looking at
how quickly patients are discharged and whether they’re readmitted. They can also
study which medicines work best, spot previously unrecognized disease patterns,
identify at-risk patients and even review the performance of individual physicians.”

2 Whiting, Rick. “Analytics Move to the Clinic,” Information Week, Healthcare Enterprise special edition
(Spring 2004)

7
Analytics in Healthcare

An effective framework for this purpose would seamlessly integrate these

fundamental components:

• A centralized data repository that synthesizes data from currently incompatible

data silos on any platform and format

• Sophisticated extract, transform and load (ETL) processes that maintain data
quality, so that you can have faith in the accuracy of research based on that data

• Healthcare-specific analytics that enable non-statisticians to surface meaningful

intelligence from vast amounts of information about patients, populations,
providers, procedures and risks

• Predictive analytics to deliver more accurate research forecasts, evidence-based ■ Freeform analysis can uncover
treatment protocols and improved patient outcomes unexpected patterns or rules
that researchers, clinicians and
• Query, reporting and visualization tools that give various types of users the
highest quality of information, where and when needed, via multiple platforms managers can use to support
and channels evidencebased change.

SAS solutions for healthcare providers

With SAS solutions for healthcare providers, you can get all of these essential
ingredients from one vendor – including analytic models and reporting templates.

All SAS solutions for healthcare are built on the SAS Intelligence Platform, a
technology framework that extends intelligence to all operating units across the
organization and to all types of users, from research scientists to clinicians to
administrators and regulators.

The use of consistent metadata (the data that describes how data elements are
derived, used and managed) allows data to flow seamlessly across all diverse
platforms and applications – as well as other SAS healthcare solutions, which all
use the same SAS foundation.

SAS analytic solutions come in all levels of sophistication, from desktop

applications for clinical quality teams to sophisticated enterprise wide performance
management systems.

8
 Analytics in Healthcare

The solution supports structured and unstructured analysis:

• Structured analysis enables you to evaluate a parameter or hypothesis, such

as analyzing reimbursement trends for a diagnostic code or length of stay for a
particular condition or procedure.

• Unstructured analysis reveals hidden patterns in masses of data, without having a

predetermined idea or hypothesis about what the pattern may be – for example,
predicting the response to changes in treatment protocols for ventilator patients.

When you can slice-and-dice, drill down or roll up through various dimensions in a
database – applying sophisticated models and algorithms in the process – you can
uncover very important patterns and best practices.

For years, healthcare organizations have used SAS analytic solutions to achieve
business goals related to cost control, revenue generation and strategic performance
management.

Now that EMR/EHR systems are becoming more commonplace – and organizations
are capturing a wealth of new clinical data – SAS analytics are helping to
revolutionize the practice of medicine, not just medical practices.

For example:

• In a medical center, SAS data mining and analytic tools are used to determine
whether certain patterns of physician prescribing will optimize outcomes for
patients undergoing open heart surgery, taking patient risk factors into account.

• Across a three-facility health system, SAS analytic tools are used to explore
clinical outcomes and risk tolerances to improve the overall quality of patient care.

• A community hospital team used SAS to track the long-term care and
rehabilitation of spinal surgery patients, to demonstrate to Medicare that although
upfront costs were high, cumulative costs were lower than other hospitals and
patients returned to productive lives in the community.

• In a health studies institute, SAS analytic tools are used to determine whether
certain clinical pathways optimize patient outcomes, in order to achieve
best practices.

• Teams in a large healthcare system use SAS for more than 100 projects a year –
to assess patient mental states, optimize operating room use by specialty and
examine outcomes relative to the use of certain drugs, medical devices and
protocols.

9
 Analytics in Healthcare

The data-based evolution of medicine

For all its Star Trek technological marvels – CAT scans, MRIs, LASIKs and so on –
the practice of medicine still operates on a somewhat traditional model. Besides still
being highly manual in record-keeping, the practice tends to be reactive: focusing on
the treatment of an existing condition.

New data resources are improving the efficiency, cost, accuracy and outcomes of
medical treatments, but the future holds even greater potential. Predictive analysis of
diverse data promises to revolutionize the essential model of healthcare delivery. With
predictive insights based on vast amounts of clinical and research data, the focus
can shift more toward proactive and preventive care.

Sophisticated data analysis is already fueling three exciting trends in medicine:

• Evidence-based medicine focuses on the conscientious, explicit and judicious

use of current best evidence in making decisions about patient care – integrating ■ SAS leverages the investments you
clinical expertise with the best available research results.
have already made (or are about
• Protocol-based medicine brings the results of research to the bedside in to make) in EMR/EHR systems and
the form of best practices that are continuously updated based on analytic applications by adding a unique
assessment of patient populations.
layer of intelligence you can’t get
• Personalized medicine acknowledges the uniqueness of individuals and provides anywhere else.
the right treatment in the right format to the right individual at the right time. This
approach blends genotypic data, medical images, environmental data, genetic
profiles, molecular and genetic research efforts and targeted pharmaceuticals with
historical clinical data to lead to personalized diagnosis and treatment.

Evidence-based medicine focuses on how medicine is practiced. Protocol-driven

medicine describes a methodology for advancing research through clinical practice.
Personalized medicine describes an ideal (but achievable) future state of customized
care. With EMR/EHR, coupled with predictive analytics, the healthcare industry is
advancing forward into all three ideals.

10
Analytics in Healthcare

11
SAS Institute Inc. World Headquarters   +1 919 677 8000
To contact your local SAS office, please visit: www.sas.com/offices
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA
and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies.
Copyright © 2009, SAS Institute Inc. All rights reserved. 102465_544428.0809